EP 3640856 A1 20200422 - A METHOD, APPARATUS AND COMPUTER PROGRAM TO CARRY OUT A TRAINING PROCEDURE IN A CONVOLUTIONAL NEURAL NETWORK
Title (en)
A METHOD, APPARATUS AND COMPUTER PROGRAM TO CARRY OUT A TRAINING PROCEDURE IN A CONVOLUTIONAL NEURAL NETWORK
Title (de)
VERFAHREN, VORRICHTUNG UND COMPUTERPROGRAMM ZUR DURCHFÜHRUNG EINES TRAININGSVERFAHRENS IN EINEM NEURONALEN FALTUNGSNETZWERK
Title (fr)
PROCÉDÉ, APPAREIL ET PROGRAMME INFORMATIQUE POUR METTRE EN UVRE UNE PROCÉDURE D'APPRENTISSAGE DANS UN RÉSEAU NEURONAL CONVOLUTIONNEL
Publication
Application
Priority
EP 18201443 A 20181019
Abstract (en)
A computer-implemented method comprises, in a computing network comprising a plurality of X nodes having processors and memory dividing neurons of a Convolutional Neural Network, CNN, between the nodes 1 to X; allocating a mini-batch of input data to each of the nodes; splitting the mini-batches into a number of sections X corresponding and equal to the number of nodes; at each node retaining the section of the mini-batch which has the same number as the node and sending the other sections of the mini-batch to their corresponding nodes; collating the mini-batch sections at each node into a single matrix and multiplying the collated matrix by the neurons at that node to provide output data having one section of output data per mini-batch; at each node sending the output data sections corresponding to the other nodes back to the corresponding nodes and combining the output data in each node so that each node has output data for its entire mini-batch.
IPC 8 full level
G06N 3/04 (2006.01); G06N 3/063 (2006.01)
CPC (source: EP US)
G06F 9/5027 (2013.01 - US); G06N 3/045 (2023.01 - EP); G06N 3/0464 (2023.01 - EP); G06N 3/063 (2013.01 - US); G06N 3/084 (2013.01 - EP); G06N 3/09 (2023.01 - EP); G06N 3/098 (2023.01 - EP); G06T 1/20 (2013.01 - US)
Citation (applicant)
- CAFFE - Y. JIA; E. SHELHAMER; J. DONAHUE; S. KARAYEV; J. LONG; R. GIRSHICK; S. GUADARRAMA; T. DARRELL: "Caffe: Convolutional Architecture for Fast Feature Embedding", ARXIV PREPRINT ARXIV:1408.5093, 2014
- TENSORFLOW - M. ABADI; A. AGARWAL; P. BARHAM, LARGE-SCALE MACHINE LEARNING ON HETEROGENEOUS DISTRIBUTED SYSTEMS, 2015
- MXNET - T. CHEN; M. LI; Y. LI; M. LIN; N. WANG; M. WANG; T. XIAO; B. XU; C. ZHANG; Z. ZHANG: "MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems", NEURAL INFORMATION PROCESSING SYSTEMS, WORKSHOP ON MACHINE LEARNING SYSTEMS, 2015
- A. KRIZHEVSKY: "One weird trick for parallelizing convolutional neural networks", CORR, 2014
- ZOU - Y. ZOU; X. JIN; Y. L; Z. G. E. WANG; B. XIAO: "Mariana: Tencent deep learning platform and its applications", PROCEEDINGS OF THE VLDB ENDOWMENT, vol. 7, no. 13, 2014, pages 1772 - 1777
Citation (search report)
- [I] US 2017148433 A1 20170525 - CATANZARO BRYAN [US], et al
- [A] US 9646243 B1 20170509 - GOKMEN TAYFUN [US]
- [A] US 2018032865 A1 20180201 - NISHIMURA HIROKI [JP], et al
- [A] US 2016171974 A1 20160616 - HANNUN AWNI [US], et al
Designated contracting state (EPC)
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR
Designated extension state (EPC)
BA ME
DOCDB simple family (publication)
EP 3640856 A1 20200422; JP 2020068016 A 20200430; JP 7334551 B2 20230829; US 11687763 B2 20230627; US 2020125933 A1 20200423
DOCDB simple family (application)
EP 18201443 A 20181019; JP 2019166073 A 20190912; US 201916600002 A 20191011